[
 {"from": "prereg:f71f0995d50a1e35c4ea4c9d639d9974f265440ef74f91c461a8964926977969", "kind": "changed", "detail": "Added code/summarize.py, which reads the registered outputs results/R1.json and results/R2.json and writes results/S.json: the same values rounded for display (p-values to two significant digits, rejection rates to three decimals, statistics to two), the minimum and maximum across the five records, and counts of records below the registered thresholds. code/run calls it after code/analyze.py. The registered analysis code (analyze.py, fr.py, tmax.py), parameters and seeds are unchanged, so R1.json and R2.json are exactly what the registered plan produces; the claims cite S.json for readability.", "claims": ["C1", "C2", "C3", "C4"]}
]